🏁 1/10th-scale autonomous racing

I was a core member of CMU’s F1TENTH team. Based on the F1TENTH platform, we built a 1:10 scaled autonomous race car and developed autonomy software including SLAM, state estimation, raceline optimization, and tracking control. It is able to perform head-to-head racing and reach maximum velocity of 20 km/h. I participated in the competition in San Antonio within the Cyber-Physical Systems and Internet-of-Things Week in May 2023 and the team advanced to semi-final.

During my involvement, I worked on integrated planning and control from simulation to hardware. Things I used for the project

  • SLAM: ROS 2 slam_toolbox
  • State estimation: particle filter
  • Raceline optimization: TUM software
  • Control: gap-following, wall-following, PID, pure-pursuit, MPC.
Autonomous driving during the race.
Race map.

Note: we ended up using simple methods for the race.